Input-Output Feedback Linearization Control for a PWR Nuclear Power Plant

This study proposes a feedback linearization-based control using a dynamic neural network to control a pressurized water-type nuclear power plant. The nonlinear plant model adopted in this study is characterized by five inputs, five outputs and, 38 state variables. The model is linearized through dynamic neural network-based system identification and feedback linearization. The proportional-integral-derivative (PID) controller is subsequently applied to the linearized process. The effectiveness of the proposed approach is demonstrated by simulations on different subsystems of a pressurized water reactor nuclear power plant model. Simulation results show that the proposed strategy offers good performance and is capable of effectively tracking the reference under disturbances.